Acoustic emission and force sensor fusion for monitoring the cutting process

Abstract An acoustic emission and force-based sensor fusion system involving pattern recognition analysis has been used to detect tool breakage, chip form and a threshold level of tool flank wear in turning. When normalized with the resultant force, the force components in the cutting, radial and feed directions were found to be highly sensitive to variables such as feedrate, material hardness, tool coating and tool wear, depth of cut, and speed in fractional factorial experiments. A three-dimensional analytical force model was extended to include the effect of flank wear in order to interpret the experimental findings. Subsequently, using an empirical bilinear relationship between the machining variables and forces, a filter was designed to eliminate the variable effects such that pattern recognition of tool failure under varying conditions was feasible. Results of the sensor fusion approach involving testing the system with the same data used in designing it when using AE and force signals indicate a 94% accuracy for sensing tool wear alone, whereas using only AE for detecting chip form and tool breakage indicate a 99 and 96% accuracy respectively.